mlsanityData to 7 Oct 2026

The Clever Hans Effect in Unsupervised Learning

arXiv 2408.08041 ↗First posted 15 Aug 2024cs.LG cs.AI stat.MLcited by 24

Not on the map. Why?

Abstract (arXiv)

Unsupervised learning has become an essential building block of AI systems. The representations it produces, e.g. in foundation models, are critical to a wide variety of downstream applications. It is therefore important to carefully examine unsupervised models to ensure not only that they produce accurate predictions, but also that these predictions are not "right for the wrong reasons", the so-called Clever Hans (CH) effect. Using specially developed Explainable AI techniques, we show for the first time that CH effects are widespread in unsupervised learning. Our empirical findings are enriched by theoretical insights, which interestingly point to inductive biases in the unsupervised learning machine as a primary source of CH effects. Overall, our work sheds light on unexplored risks associated with practical applications of unsupervised learning and suggests ways to make unsupervised learning more robust.

Key facts. Cited by 24 papers in the map (data through 7 Oct 2026). Builds on 43 of its 96 references that are in the map. First posted on arXiv 15 Aug 2024.

Builds on

43 of 96 references.

References · 43

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  4. 2020Language Models are Few-Shot Learnerscited by 32,199
  5. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  6. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,617
  7. 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
  8. 2020Shortcut Learning in Deep Neural Networkscited by 1,928
  9. 2021Barlow Twins: Self-Supervised Learning via Redundancy Reductioncited by 1,818
  10. 2019Explainable Artificial Intelligence (XAI): Concepts, Taxonomies, Opportunities and Challenges toward Responsible AIcited by 1,543
  11. 2020Big Self-Supervised Models are Strong Semi-Supervised Learnerscited by 1,375
  12. 2017ChestX-ray8: Hospital-scale Chest X-ray Database and Benchmarks on Weakly-Supervised Classification and Localization of Common Thorax Diseasescited by 969
31 more
  1. 2021Towards Total Recall in Industrial Anomaly Detectioncited by 772
  2. 2020A Unifying Review of Deep and Shallow Anomaly Detectioncited by 423
  3. 2022Last Layer Re-Training is Sufficient for Robustness to Spurious Correlationscited by 379
  4. 2019Unmasking Clever Hans Predictors and Assessing What Machines Really Learncited by 359
  5. 2017Methods for Interpreting and Understanding Deep Neural Networkscited by 342
  6. 2021Big Self-Supervised Models Advance Medical Image Classificationcited by 320
  7. 2020COVID-Net: A Tailored Deep Convolutional Neural Network Design for Detection of COVID-19 Cases from Chest X-Ray Imagescited by 264
  8. 2019Probing Neural Network Comprehension of Natural Language Argumentscited by 212
  9. 2020Deep Learning for Anomaly Detection: A Reviewcited by 159
  10. 2023EfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latenciescited by 157
  11. 2020Making deep neural networks right for the right scientific reasons by interacting with their explanationscited by 141
  12. 2020Frequency Bias in Neural Networks for Input of Non-Uniform Densitycited by 140
  13. 2020Higher-Order Explanations of Graph Neural Networks via Relevant Walkscited by 138
  14. 2020Intriguing Properties of Contrastive Lossescited by 129
  15. 2020Modeling the Distribution of Normal Data in Pre-Trained Deep Features for Anomaly Detectioncited by 115
  16. 2021Can contrastive learning avoid shortcut solutions?cited by 114
  17. 2020COVID-19 Image Data Collection: Prospective Predictions Are the Futurecited by 100
  18. 2022XAI for Transformers: Better Explanations through Conservative Propagationcited by 100
  19. 2022Human alignment of neural network representationscited by 74
  20. 2024RudolfV: A Foundation Model by Pathologists for Pathologistscited by 40
  21. 2021Spectral Bias in Practice: The Role of Function Frequency in Generalizationcited by 36
  22. 2020Building and Interpreting Deep Similarity Modelscited by 34
  23. 2024MambaLRP: Explaining Selective State Space Sequence Modelscited by 31
  24. 2022Disentangled Explanations of Neural Network Predictions by Finding Relevant Subspacescited by 25
  25. 2018Towards Explaining Anomalies: A Deep Taylor Decomposition of One-Class Modelscited by 24
  26. 2019Explaining and Interpreting LSTMscited by 23
  27. 2023Explainable AI for Time Series via Virtual Inspection Layerscited by 23
  28. 2021Towards Fine-grained Visual Representations by Combining Contrastive Learning with Image Reconstruction and Attention-weighted Poolingcited by 13
  29. 2020Addressing Feature Suppression in Unsupervised Visual Representationscited by 11
  30. 2023Preemptively Pruning Clever-Hans Strategies in Deep Neural Networkscited by 9
  31. 2023Insightful analysis of historical sources at scales beyond human capabilities using unsupervised Machine Learning and XAIcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 3; 2025: 9; 2026: 12.

24 papers.

Sort:

By area:

  1. 2025Mechanistic understanding and validation of large AI models with SemanticLenscited by 29
  2. 2024Do Histopathological Foundation Models Eliminate Batch Effects? A Comparative Studycited by 17
  3. 2024Towards Symbolic XAI -- Explanation Through Human Understandable Logical Relationships Between Featurescited by 7
  4. 2025Towards Desiderata-Driven Design of Visual Counterfactual Explainerscited by 7
  5. 2024Opportunities and limitations of explaining quantum machine learningcited by 6
  6. 2025MeDi: Metadata-Guided Diffusion Models for Mitigating Biases in Tumor Classificationcited by 6
  7. 2025Mitigating Clever Hans Strategies in Image Classifiers through Generating Counterexamplescited by 6
  8. 2026Atlas 2 -- Foundation models for clinical deploymentcited by 6
  9. 2025Position: We Need An Algorithmic Understanding of Generative AIcited by 5
  10. 2026Visual Disentangled Diffusion Autoencoders: Scalable Counterfactual Generation for Foundation Modelscited by 4
  11. 2025Model Science: getting serious about verification, explanation and control of AI systemscited by 3
  12. 2025Wasserstein Distances Made Explainable: Insights Into Dataset Shifts and Transport Phenomenacited by 2
12 more
  1. 2026Distilling Lightweight Domain Experts from Large ML Models by Identifying Relevant Subspacescited by 2
  2. 2026Where Not to Learn: Prior-Aligned Training with Subset-based Attribution Constraintscited by 2
  3. 2026Beyond Attention Heatmaps: How to Get Better Explanations for Multiple Instance Learning Models in Histopathologycited by 2
  4. 2026Normalized Relevance Measure as a Unifying Framework to Explain Neural Network Latent Structurescited by 2
  5. 2025Reading Smiles: Proxy Bias in Foundation Models for Facial Emotion Recognitioncited by 1
  6. 2025Fast and Accurate Explanations of Distance-Based Classifiers by Uncovering Latent Explanatory Structurescited by 0
  7. 2026Stable but Wrong: When Learning Stabilizes Away from the Truthcited by 0
  8. 2026Symb-xMIL: Symbolic Explanations for Multiple Instance Learning in Digital Pathologycited by 0
  9. 2026Explaining Temporal Graph Neural Networks via Feature-induced Information Flowcited by 0
  10. 2026Position: Reasoning is a Learnable Rule-Based Processcited by 0
  11. 2026HERO: Histology Encoder for Robust Representation in Oncologycited by 0
  12. 2026Towards Fast and Disentangled Counterfactuals for Visual Foundation Modelscited by 0